Late perfusion defects after pulmonary embolism and long-term mortality: a weighted cohort study

Residual pulmonary vascular obstruction (RPVO) after pulmonary embolism (PE) is common and has been associated with adverse outcomes, although its independent prognostic significance remains uncertain.

To investigate whether perfusion defects are independently associated with long-term mortality after PE, and secondarily to assess their relationship with dyspnoea severity.

In this nationwide cohort study, Swedish patients with PE in 2005 who survived to 2007 underwent structured follow-up including symptom assessment and clinical evaluation. Ventilation-perfusion (V/Q) scintigraphy was recommended based on clinical judgement. V/Q scintigraphy findings were classified as normal or perfusion defects. To account for selection bias into scintigraphy, inverse probability weighting (IPW) was applied. Survival was analysed using IPW-weighted Cox regression with delayed entry at scintigraphy.

Among 911 patients eligible for scintigraphy, 523 underwent examination; 513 with normal findings or perfusion defects were included in the analyses. Median time from PE to scintigraphy was 6.3 years, with 8.8 years of follow-up thereafter. Perfusion defects were present in 221 patients (43%). Dyspnoea severity was not associated with perfusion defects in unadjusted or adjusted analyses. In unadjusted analyses, perfusion defects were associated with lower survival (log-rank p = 0.049). However, this association was attenuated after adjustment and was not independently associated with mortality in IPW-weighted Cox regression (hazard ratio 1.28, 95% confidence interval 0.94-1.74).

Perfusion defects were associated with mortality in unadjusted analyses but not after adjustment for age and comorbidities, suggesting that their apparent prognostic impact largely reflects underlying patient characteristics rather than an independent effect.